In 1956, the founding fathers of artificial intelligence convened under the assumption that “… every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.” (Dartmouth Artificial Intelligence Project Proposal)
This idea was dominant until the mid-1980s. More recently there has been increasing interest in “embodiment”- the notion that intelligence (behaviour) is not only a matter of computation, but requires a body, a complete organism that interacts with the real world. As a consequence, many researchers shifted their attention from computers to robots, from screen interfaces to tangible media, from virtual reality to mixed reality.
In this presentation, the far-reaching and often surprising implications of embodiment will be explored. There are deeper and more important consequences beyond the trivial meaning of “intelligence requires a body”. This paper is concerned about modelling physical agent behaviours without a complete computational model.
The discussion will approach the results of “outsourcing” behavioural models into physical constructions where the apparent actions emerge from the interaction of the agent within the physical world with minimal computational efforts and representational models.
Being situated, theses agents can also empower intangible contributions from the cultural context, the suspension of disbelief and the attribution of intention towards any outside physical objects acting upon the world. The paper will analyze, from the audience standpoint, where our perception departs from the simple function (mechanical or programmed) towards the intention (self motivated complex agent).
The paper will present robots that express emotions mainly due to their intrinsic materials and the very complex dynamics of their structure in motion. Furthermore, there will be a discussion about on how to carry these results into the realm of mixed realities and man-machine interactions.
Louis-Philippe Demers Biography
Full text (PDF)